kopano-server shutdown after rsyslog running

goldingerit

hi forum

we have an issue with our kopano server on debian with rsyslog.

on last sunday at 06.26am service “kopano-server” crashed.

Syslog:

May 21 06:26:11 System01 kernel: [896930.101525] kopano-server[26933]: segfault at 726576726553 ip 00007fd2ab7a946e sp 00007fd2a5187f30 error 4 in libtcmalloc_minimal.so.4.1.0[7fd2ab78d000+29000]
May 21 06:26:11 System01 rsyslogd: [origin software="rsyslogd" swVersion="5.8.11" x-pid="16075" x-info="http://www.rsyslog.com"] rsyslogd was HUPed
May 21 06:26:12 System01 rsyslogd: [origin software="rsyslogd" swVersion="5.8.11" x-pid="16075" x-info="http://www.rsyslog.com"] rsyslogd was HUPed
May 21 06:26:12 System01 kernel: imklog 5.8.11, log source = /proc/kmsg started.
May 21 06:26:12 System01 rsyslogd: [origin software="rsyslogd" swVersion="5.8.11" x-pid="372" x-info="http://www.rsyslog.com"] start

kopano-server.log:

Sun May 21 06:26:11 2017: [warning] Log connection was reset

debian version:
7.11

rsyslog version:
5.8.11

kopano-server version:
8.4.0.347

we have the same issue on multiple servers, but some are ok. all servers have same versions (debian, rsyslog and kopano-server)

thanks in advance for any help.

regards,
daniel

fbartels

Can you reproduce the error with a higher loglevel?

goldingerit

hi felix

with debug log_level (6) i get this error message in kopano/server.log:
Sun May 28 06:50:32 2017: [warning] Log connection was reset
Sun May 28 06:50:32 2017: [crit ] Config error: Option ‘ldap_group_type_attribute_value’ cannot be empty!
Sun May 28 06:50:32 2017: [crit ] Config error: Option ‘ldap_user_type_attribute_value’ cannot be empty!

we have
in /etc/kopano/ldap.cfg:
…
!include /usr/share/kopano/ldap.active-directory.cfg
…

and in /usr/share/kopano/ldap.active-directory.cfg:
…
ldap_user_type_attribute_value = user
ldap_group_type_attribute_value = group
…

thanks for your help.

regards,
daniel

fbartels

are you really sure the process has shutdown? your logging has nothing that indicates this.

goldingerit

yes, it has shutdown.
same error in syslog:

May 28 06:50:32 System01 kernel: [326160.209965] kopano-server[11964]: segfault at 0 ip 00007f2c13cb32cb sp 00007f2c0d697060 error 4 in libtcmalloc_minimal.so.4.1.0[7f2c13c9c000+29000]

and /etc/init.d/kopano-server status is:
kopano-server is not running.

but manually starting the service with /etc/init.d/kopano-server start works fine.

maybe there is a problem in /etc/logrotate.d/kopano?

/var/log/kopano/archiver.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
}

/var/log/kopano/dagent.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-dagent 2> /dev/null || true
        endscript
}

/var/log/kopano/gateway.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-gateway 2> /dev/null || true
        endscript
}

/var/log/kopano/ical.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-ical 2> /dev/null || true
        endscript
}

/var/log/kopano/search.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-search 2> /dev/null || true
        endscript
}

/var/log/kopano/monitor.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-monitor 2> /dev/null || true
        endscript
}

/var/log/kopano/server.log /var/log/kopano/audit.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-server 2> /dev/null || true
        endscript
}

/var/log/kopano/spooler.log {
        weekly
        missingok
        rotate 52
        compress
        delaycompress
        notifempty
        postrotate
                        killall -HUP kopano-spooler 2> /dev/null || true
        endscript
}

regards,
daniel

fbartels

@goldingerit ah, I did miss the segfault in syslog for kopano-server. Does the segfault happen when logrotating?

goldingerit

Yes, it happens at the same time.